This study evaluated the efficacy of computed tomography (CT)-based three-dimensional (3D) dose-volume parameters of the rectum as predictor for late rectal complication (LRC) in cervical cancer patients treated with radiotherapy alone. Eighty-four patients treated with a combination of external radiotherapy and high-dose-rate intracavitary brachytherapy between January 2000 and December 2004 were retrospectively analyzed. Brachytherapy was prescribed with standard 2D planning. Patients underwent pelvic CT at brachytherapy. The external rectal wall was contoured on the CT images, and the minimum doses delivered to 0.1cc, 1cc, and 2cc of the most irradiated rectal volumes were calculated with dose-volume histograms. The International Commission of Radiation Units and Measurements (ICRU) rectal point dose was also calculated by conventional method. Total dose (external radiotherapy plus brachytherapy) to the rectum was transformed to the biologically equivalent dose in 2-Gy fractions with alpha/beta of 3 Gy (D(0.1cc), D(1cc), D(2cc) and D(ICRU)). The relationships between these dosimetric parameters and the incidence of LRC were analyzed. The 5-year overall actuarial rate of LRC was 26.4%. The values of D(0.1cc), D(1cc), and D(2cc) were significantly higher in patients with LRC than in those without (p < 0.001), but the difference in the values of D(ICRU) was not statistically significant (p = 0.10). The rate of LRC increased significantly with increasing D(0.1cc), D(1cc), and D(2cc) (p = 0.001). However, no positive dose-response relationship was observed between D(ICRU) and the rate of LRC (p = 0.42). The present study has suggested that CT-based 3D dose-volume parameters of the rectum may be effective for predicting LRC.